Title: Generalized covariance estimator
Authors: Joann Jasiak - York University (Canada) [presenting]
Christian Gourieroux - University of Toronto and CREST (Canada)
Abstract: A class of semi-parametric dynamic models is considered with strong white noise errors including the standard Vector Autoregressive (VAR) model, the nonfundamental structural VAR model, the mixed causal-noncausal models and nonlinear dynamic models such as the (multivariate) ARCH-M model. For this class of models, we propose the Generalized Covariance (GCov) estimator, which is obtained by minimizing a residual-based multivariate portmanteau statistic. The GCov is a reliable alternative to the Generalized Method of Moments (GMM) providing semi-parametrically efficient estimates in one step. We derive the asymptotic properties of the GCov estimator and show that the associated residual-based portmanteau statistic is asymptotically chi-square distributed. The finite sample performance of the GCov estimator is illustrated in a simulation study. The estimator is also applied to a dynamic model of cryptocurrency prices.